Classical cancer models to date which have attempted to explain the character of cancer cells have typically described them as fast-growing and highly mutant cells. These cancer cells are hypothesized to have been produced during carcinogenesis because of a multi-step neo-Darwinian evolutionary process involving mutation-selection events at the cellular level (Fearon et al, “A Genetic Model for Colorectal Tumorigenesis”, Cell, 61:759-767 (1990); Nowell, “The Clonal Evolution of Tumor Cell Populations”, Science (Washington, D.C.), 194:23-28 (1976)).
Related to this, conventional cancer diagnoses and therapies to date have attempted to selectively detect and eradicate neoplastic cells that are largely fast-growing and mutant. For example, conventional cancer chemotherapies (e.g., alkylating agents such as cyclophosphamide, anti-metabolites such as 5-Fluorouracil, plant alkaloids such as vincristine) in a similar manner to conventional irradiation therapies both exert their toxic effects on cancer cells by interfering with numerous cellular mechanisms involved in cell growth and DNA replication. Other less commonly used experimental cancer therapies have included use of immunotherapies wherein the administration of therapeutic agents which selectively bind mutant tumor antigens on fast-growing cancer cells (e.g., monoclonal antibodies) have been attached to therapeutic moieties such as toxins, radionuclides or chemotherapeutic agents for the purpose of eradicating fast-growing mutant cancer cells. Similarly, newer experimental gene-directed therapies have attempted to exploit certain cancer-related mutations by correcting or replacing such defects by, e.g., inserting wildtype versions of such mutant genes (e.g., p53) into cancer cells, inhibiting overactive oncogenes (e.g., ras) in tumor cells, or by using existing mutations in cancer cells as targets for other therapeutic (e.g., viral) vectors (Feng et al, “Neoplastic reversion accomplished by high efficiency adenoviral-mediated delivery of an anti-ras ribozyme”, Cancer Res., 55:2024-2028 (1995); Bischoffet al, “An adenovirus that replicates in p53-deficient human tumor cells”, Science, 274:373-376 (1997)). While all of these methods (i.e., conventional chemotherapies, irradiation, immunotherapies, and gene therapies) may, by their design, eradicate a significant proportion of a given tumor mass by destroying the large populations of highly proliferative and mutant neoplastic cells thus resulting in a clinical remission, in time the tumor may recur at the same or different site(s), thereby indicating that not all cancer cells have been eradicated by these methods. A number of reasons for tumor relapse have been offered within the conventional paradigm. These include insufficient chemotherapeutic dosage (limited by onset of significant side effects), and/or emergence of cancer clones which are resistant to therapy.
The novel model for carcinogenesis presented here (termed the OSES model) offers an alternative explanation for relapse wherein (as will be discussed in more detail) a clandestine slow-growing relatively mutationally-spared cancer stem line acts as the immortal founder line of a tumor and produces as its progeny the highly proliferative mutant cancer cell populations targeted by conventional therapies mentioned above. Accordingly, it will be shown that this cancer stem line (hypothesized to exist by the OSES model) is not targeted by conventionally-based therapies (designed to target fast-growing largely mutant cells rather than slow-growing non-mutant cells). In this manner, the untargeted cancer stem line can gradually regrow the tumor mass following standard therapy thereby leading to treatment failure and clinical relapse.
Recently much evidence has been accumulated which raises significant concerns as to the validity of classical cancer models based on the neo-Darwinian paradigm—and thus also to the efficacies of cancer therapies wholly based on this model. So that the invention may be understood, previous conventional cancer models and their inadequacies are discussed below.
There is extensive evidence that cancer results from the evolution of an increasingly “cancer-like” tissue type leading ultimately to one with malignant capability (Furth, “Conditioned and autonomous neoplasms: a review”, Cancer Res., 13:477-492 (1953); Foulds, “The natural history of cancer”, J. Chronic Dis., 8:2-37 (1958)). That mutagenesis is causally involved in the initiation of this process is a concept which has stemmed largely from demonstrations that genotoxic carcinogens can cause cellular transformation and that this “initiated” phenotype is rare, permanent, focal and heritable. (Berenblum, “Sequential aspects of chemical carcinogenesis: skin, in: F. F. Becker (ed.), Cancer: a Comprehensive Review, pp. 451-484, New York, Plenum Publishing Corp. (1982); Cohen et al, “Genetic Errors, Cell Proliferation, and Carcinogenesis”, Cancer Res., 51:6493-6505 (1991)). Moreover, that a mutant gene can bestow a transmissible cancer phenotype is evidenced by transfection and transgenic experiments as well as recent genetic linkage analyses of human familial cancers (Weinberg, “Oncogenes, Anti-oncogenes, and the Molecular Basis of Multistep Carcinogenesis”, Cancer Res., 49: 3713-3721 (1989); Tsukamoto et al, “Expression of the int-1 gene in transgenic mice is associated with mammary gland hyperplasia and adenocarcinomas in male and female mice”, Cell, 55: 619-625 (1988); Knudson, “Hereditary cancer, oncogenes, and antioncogenes”, Cancer Res., 45: 1437-1343 (1985)). It is also clear that while the onset of carcinogenesis requires an initial (acquired or inherited) mutagenic insult, subsequent alterations are also necessary for attainment of malignancy (Nowell, “The Clonal Evolution of Tumor Cell Populations.”, Science (Washington D.C.), 194: 23-28 (1976)).
This concept is supported by findings that certain cultured cells transfected with a single oncogene require additional alterations to become fully transformed. (Weinberg, “Oncogenes, Anti-oncogenes, and the Molecular Basis of Multistep Carcinogenesis”, Cancer Res., 49:3713-3721 (1989)). Moreover, epidemiological data implicating a series of rate limiting steps in the pathogenesis of human cancers has lent support to the concept that a gradual cellular progression toward increasing malignancy is driven by neo-Darwinian mutation-selection (Miller, “On the Nature of Susceptibility to Cancer”, Cancer, 46:1307-1318 (1980).
Also, the well-documented association between increasing mutational load and tumor grade has led to the seemingly most parsimonious mechanistic explanation of these data whereby mutations cause both the initiation as well as progression of neoplasia via a continuum of mutation-selection events at the cellular level (Nowell, “The Clonal Evolution of Tumor Cell Populations.”, Science (Washington D.C.), 194: 23-28 (1976); Fearon et al, “A Genetic Model for Colorectal Tumorigenesis”, Cell, 61: 759-767 (1990)). Additional support for this neo-Darwinian model derives from recent high-resolution molecular analyses of human tumor biopsy specimens that reveal a tight correlation between the appearance of certain defined genetic alterations and the transition to increasingly “cancerous-appearing” tumor regions (Fearon et al, “A Genetic Model for Colorectal Tumorigenesis”, Cell, 61: 759-767 (1990); Sidransky et al, “Clonal expansion of p53 mutant cells is associated with brain tumor progression”, Nature (Lond.), 355: 846-847 (1992); Sato et al, “Allelotype of Breast Cancer: Cumulative Allele Losses Promote Tumor Progression in Primary Breast Cancer”, Cancer Res., 50: 7184-7189 (1990)).
However, it should be noted that there is a body of cancer literature not readily accounted for by standard neo-Darwinian mutation-selection models. Namely, there are a host of independent data describing unexpectedly elevated transformation rates observed in certain carcinogen-treated cells not fully accounted for by somatic mutation alone as well as the capability of some highly malignant tumor types to differentiate or even revert to normal under certain conditions—findings also not readily explained by conventional mutation-selection models (Kennedy et al, “Timing of the steps in transformation of C3H 10T1/2 cells by X-irradiation”, Nature (Lond.), 307: 85-86 (1984); Rubin, “Cancer as a Dynamic Developmental Disorder”, Cancer Res., 45: 2935-2942 (1985); Farber et al, “Cellular Adaptation in the Origin and Development of Cancer”, Cancer Res., 51: 2751-2761 (1991)). Accordingly, alternative models to mutation-selection (e.g., invoking a role for potentially reversible non-mutational/epigenetic alterations as effectors of cellular evolution toward increasing malignancy) have been advanced by several investigators in order to explain these and other related phenomena.
For example, as alluded to, while only a minority of mouse prostate cells exposed to methylcholanthrene initially become transformed, the entire population of treated cells rear progeny with an increased propensity for transformation at subsequent cell divisions despite removal of the carcinogenic agent. Similarly, treatment of various rodent cells with other types of carcinogens, namely X-irradiation or retroviral infection, also results in transformation of progeny of initially untransformed cells at an overall rate difficult to reconcile by somatic mutagenesis alone (Farber et al, “Cellular Adaptation in the Origin and Development of Cancer”, Cancer Res., 51: 2751-2761 (1991)). In addition, Rubin has demonstrated that although only a small fraction of growth-constrained NIH3T3 cells form transformed foci, the entire population of these murine fibroblasts gives rise to clones with elevated transformation rates (Rubin, “Cellular epigenetics: Control of the size, shape, and spatial distribution of transformed foci by interactions between the transformed and nontransformed cells”, Proc. Natl. Acad. Sci. USA, 91: 6619-6623 (1994)). In a related manner, while only a minority of in vivo DMBA-treated murine skin cells become transformed, heritable phenotypic alterations are present in the entire basal skin cell layer exposed to this chemical carcinogen thereby corroborating those mentioned in vitro experiments cited in support of a widespread heritable phenotypic effect by certain carcinogens beyond that which can be attributed wholly to their mutagenic effects (Farber et al, “Cellular Adaptation in the Origin and Development of Cancer”, Cancer Res., 51: 2751-2761 (1991)).
It has also been noted, as mentioned, that a variety of cancer cell types can differentiate to varying degrees. For example, human neuroblastoma cells sprout axons and dendrites when grown as explants and murine leukemic cells differentiate into benign granulocytes and macrophages when grown in vitro. Moreover, tritiated thymidine labeling of rodent squamous cell carcinomas and skeletal muscle tumors illustrates that poorly differentiated cells within these tumors can give rise to well-differentiated squamous epithelia and multinucleated myotubes, respectively. In addition, somatic tissues of transplantable mouse teratocarcinomas have been shown to be benign differentiated progeny of a subpopulation of poorly differentiated embryonal carcinoma cells within these particular tumors. Most recently, all-trans-retinoic acid (ATRA) has been found to be efficacious in the treatment of human acute promyelocytic leukemia (APL) by inducing terminal differentiation of malignant leukocytes (Pierce et al (eds.), “Cancer: a problem of developmental biology”, New Jersey: Prentice Hall Inc. (1978); Degos et al, “All Trans-Retinoic Acid as a Differentiating Agent in the Treatment of Acute Promyelocytic Leukemia”, Blood, 85: 2643-2653 (1995)).
Accordingly, it had been suggested by some investigators that conventional cancer models may not adequately explain elevated transformation rates of certain cells types (not adequately explained by somatic mutation alone), or the differentiation capability of certain tumor cells presumably having arisen via a cascade of random genetic derangements. Based on these and other related observations, Rubin, Farber, and Pierce among others have theorized that the ability of certain cancer cells to differentiate might suggest that a cancer cell originates from potent developing/renewing cells which undergo defective morphogenesis rather than from mutant cells which undergo dedifferentiation. Moreover, it has been argued by several investigators that mechanistically such defective morphogenesis is likely to arise from a series of non-mutational (i.e., epigenetic) alterations (rather than mutations) which by definition leave the genome fairly intact and differentiation-related genes functional. Cited in additional support of this cancer theory is the observation that certain malignant cells can not only produce differentiated progeny but can also do so in a relatively orderly manner which mimics normal tissue development under specified conditions. For example, neuroblastoma growth and differentiation is regulated when placed into neurula-stage embryos, and the behavior of melanoma cells is controlled when transplanted into fetal skin (Podesta et al, “The neurula stage mouse embryo in control of neuroblastoma”, Proc. Natl. Acad. Sci. USA, 81: 7608-7611 (1984); Gerschenson et al, “Regulation of melanoma by the embryonic skin”, Proc. Natl. Acad. Sci. USA, 83: 7307-7310 (1980)). In addition, leukemic cells differentiate into normal hematopoietic tissue when injected into mouse embryos during leukocyte progenitor development and placement of malignant embryonal carcinoma cells into the embryonic milieu of a developing blastocyst results in complete reversion and differentiation of these cells into multiple tissue types leading to the formation of viable murine chimeras (Pierce et al (eds.), “Cancer: a problem of developmental biology”, New Jersey: Prentice Hall Inc. (1978); Gootwine et al, “Participation of myeloid leukaemic cells injected into embryos in hematopoietic differentiation in adult mice”, Nature (London), 299: 63-65 (1982)). In a related manner, restoration of native environmental conditions at a particular adult tissue locale (i.e., without transplantation) causes certain tumors to “behave” as a normal tissue—e.g., resolution of normal physiologic hormonal balance causes some murine endocrine tumors to regress and certain rodent sarcomas induced by implantation of an inert material into connective tissues revert when proper tissue architecture is restored (Rubin, “Cancer as a Dynamic Developmental Disorder”, Cancer Res., 45: 2935-2942 (1985)).
That altered epigenesis can lead to neoplasia is an idea which is also supported a significant literature documenting the unexpected reversibility of certain carcinogen-transformed cells in culture (Kennedy et al, “Timing of the steps in transformation of C3H 10T1/2 cells by X-irradiation”, Nature (Lond.), 307: 85-86 (1984); Rubin, “Cancer as a Dynamic Developmental Disorder”, Cancer Res., 45: 2935-2942 (1985)). This idea is also supported by demonstrations that apparently non-mutational events such as transplantation of normal rodent tissues (e.g., testis, pituitary, embryonic ectoderm) to ectopic sites can lead to their neoplastic transformation, while replacement of some of these cancer cell types back into their endogenous micro-environments results in their reversion to normalcy (Rubin, “Cancer as a Dynamic Developmental Disorder”, Cancer Res., 45:2935-2942 (1985); Farber and Rubin, “Cellular Adaptation in the Origin and Development of Cancer”, Cancer Res., 51:2751-2761 (1991)). Also, evidence that some pre-malignant lesions may be reversible during their early stages has led to the proposal by Farber that a “pre-malignant” phenotype may be no more than a programmed adaptive (epigenetic) cellular response rather than an aberrant product of random permanent mutational events (Farber, “The Multistep Nature of Cancer Development”, Cancer Res., 44:4217-4223 (1984); Farber and Rubin, “Cellular Adaptation in the Origin and Development of Cancer”, Cancer Res., 51:2751-2761 (1991)).
Accordingly, Rubin, a major proponent of the epigenetic model for cancer, had theorized that the evolution toward malignancy might be due to the progressive selection (i.e., evolution) of increasingly “cancer-like” cells undergoing advantageous epigenetic fluctuations. Moreover, Rubin, as well as Prehn, have offered that the presence of mutational alterations in tumors may not always have a causal role in tumor progression, but rather may be the result of genomic instability associated with the malignant phenotype thereby representing a cancer-related “epiphenomenon”. (Rubin, “Cancer as a Dynamic Developmental Disorder”, Cancer Res., 45:2935-2942 (1985); Prehn, “Cancers Begets Mutations versus Mutations Beget Cancers”, Cancer Res., 54:5296-5300 (1994).
However, considering increasing current molecular evidence in support of a causal role for mutagenesis in human cancer, it is no wonder that attempts have been made not only to consider alternative explanations for the action of epigenesis in carcinogenesis (i.e., other than aberrant morphogenesis which has been linked to the unpopular notion of a non-causal epiphenomenal role for mutations in cancer), but also more radically to regard epigenesis-related data in general as largely anecdotal. Notwithstanding, documentation that a variety of epigenetic alterations (such as changes in DNA methylation and genomic imprinting) are present in human biopsy specimens of a number of different tumor types has undoubtedly helped to resurrect epigenesis as a real cancer-related entity rather than an experimental quirk (Goelz et al, “Hypomethylation of DNA from benign and malignant human colon neoplasms”, Science (Washington, D.C.), 228: 187-190 (1985); Feinberg, “Genomic imprinting and gene activation in human cancer”, Nature Genet., 2: 110-113 (1993)). Accordingly, current cancer models largely of the mutation-selection variety now regularly include aspects of altered epigenesis into a more general neo-Darwinian paradigm wherein both mutagenesis and epigenesis contribute to cancer evolution by acting as independent effectors of cell variability (Vogelstein et al, “The multistep nature of cancer”, Trends Genet., 9: 138-141 (1993)). Thus, in contrast to past models of epigenesis (which have regarded mutations as non-causal), this combined paradigm is able to maintain the well documented causal role for mutagenesis in carcinogenesis.
However, it should be noted that while combining mutagenesis and epigenesis into one general paradigm has the obvious appeasing benefits that accompany compromise, it is still not obvious how such a combined paradigm is any more apt than pure mutagenic models to explain instances of tumor cell regulation, differentiation, or regression if it still invokes rare irreversible (genomic) derangements as major effectors of cell variability. In reference to that model which best accounts for these enigmatic data (i.e., the depiction of cancer as an aberrancy in tissue morphogenesis), it is clear that its current state of relative obscurity can be largely attributed to its unfortunate coupling to the ill-fated notion that cancer-related mutations are mere epiphenomena—a stance clearly at odds with recent data. However, by dismissing this entire model because of an overzealous error in deducing its consequences, is it possible that some have effectively “thrown out the baby with the bathwater”, so to speak? Alternatively, if one uncouples the concept of aberrant morphogenesis from any particular stance as to the causality of mutagenesis in carcinogenesis, cancer could then be viewed in a new light as an epigenetic defect in tissue morphogenesis but one which could also, in a seemingly contradictory manner, be abetted by mutations. For example, mutagenesis could cause cancer, not via standard stepwise mutation-selection, but rather by triggering an actively developing/renewing adult tissue to undergo a largely epigenetic-driven aberrant morphogenetic program. Moreover, subsequent to the birth of a cancer cell in this manner, mutagenesis could then act in another novel non-neo-Darwinian manner by blocking cancer cell reversion (i.e., rather than by promoting progression of pre-cancerous intermediates). Such a model would maintain the causal role of mutagenesis in carcinogenesis while more readily accounting for the epigenetic nature of certain cancers than does the current combined neo-Darwinian model. The preceding scenario is the basis for the OSES model for carcinogenesis, the specific mechanisms of which will be discussed in more detail.
Accordingly, this report will contend that upon closer analysis of the cancer literature: 1) the data normally cited in favor of mutation-selection are not exclusive to the conventional paradigm but rather are also consistent with an alternative and novel non-neo-Darwinian model (termed the OSES model), and that 2) this novel OSES model may have an advantage in being more able than the conventional paradigm to account for certain past as well as more recently described enigmatic cancer-related phenomena.
Thus, the present invention is based on a novel and improved model for carcinogenesis which incorporates and indeed reconciles the presence of the seemingly conflicting processes of epigenesis and mutagenesis that occur during carcinogenesis. Moreover, based on this novel model of carcinogenesis (i.e., the OSES model), the present invention further provides novel and improved methods for the diagnosis and treatment of cancer. These novel methods should alleviate and potentially prevent problems associated with those of conventional cancer diagnosis and therapy, in particular, the need for methods which provide for much earlier cancer diagnosis than is currently available, as well as the need for more successful initial therapies for cancer that are not as susceptible to tumor relapse as are conventional treatment regimens.